Grok 4.5 Release: Cheaper Than GPT and Claude With Top Quality

Unexpectedly successful release: Grok 4.5

It didn’t just reach the same level as Claude and OpenAI’s GPT models — it also turned out to be noticeably cheaper. More efficient in token usage while simultaneously delivering outstanding results — that combination truly deserves recognition. Bravo, Musk

The main advantage is obvious: it increases competitive pressure on OpenAI and Anthropic, forcing them either to lower prices or at least make their pricing more attractive, since xAI now has a model that outperforms them in terms of the overall set of characteristics.

It’s a damn cool price-to-quality ratio. Opus 4.8-level quality, but with a price tag like Sonnet 5.

Grok 4.5: $2 for input tokens / $6 for output tokens
GPT-5.6: $5 for input / $30 for output
Opus 4.8: $5 for input / $25 for output

Looks like the Cursor deal fully paid off.